Bootstrap

PerlMagick, Perl API

InstallationOverviewExample ScriptRead or Write an ImageManipulate an ImageSet an Image AttributeGet an Image AttributeCompare an Image to its ReconstructionCreate an Image MontageWorking with BlobsDirect-access to Image PixelsMiscellaneous MethodsHandling ExceptionsConstant

PerlMagick is an objected-oriented Perl interface to ImageMagick. Use the module to read, manipulate, or write an image or image sequence from within a Perl script. This makes it very suitable for Web CGI scripts. You must have ImageMagick 6.5.5 or above and Perl version 5.005_02 or greater installed on your system for PerlMagick to build properly.

There are a number of useful scripts available to show you the value of PerlMagick. You can do Web based image manipulation and conversion with MagickStudio, or use L-systems to create images of plants using mathematical constructs, and finally navigate through collections of thumbnail images and select the image to view with the WebMagick Image Navigator.

You can try PerlMagick from your Web browser at the ImageMagick Studio. Or, you can see examples of select PerlMagick functions.

Installation

UNIX

Is PerlMagick available from your system RPM repository? For example, on our CentOS system, we install PerlMagick thusly:

yum install ImageMagick-perl

If not, you must install PerlMagick from the ImageMagick source distribution. Download the latest source release.

Unpack the distribution with this command:

tar xvzf ImageMagick.tar.gz

Next configure and compile ImageMagick:


    $
    cd ImageMagick-7.1.1
    $ ./configure -with-perl
    $ make

If ImageMagick / PerlMagick configured and compiled without complaint, you are ready to install it on your system. Administrator privileges are required to install. To install, type

sudo make install

You may need to configure the dynamic linker run-time bindings:

sudo ldconfig /usr/local/lib

Finally, verify the PerlMagick install worked properly, type

perl -MImage::Magick -le 'print Image::Magick->QuantumDepth'

Congratulations, you have a working ImageMagick distribution and you are ready to use PerlMagick to convert, compose, or edit your images.

Windows XP / Windows 2000

ImageMagick must already be installed on your system. Also, the ImageMagick source distribution for Windows 2000 is required. You must also have the nmake from the Visual C++ or J++ development environment. Copy \bin\IMagick.dll and \bin\X11.dll to a directory in your dynamic load path such as c:\perl\site\5.00502.

Next, type

cd PerlMagick
perl Makefile.nt
nmake
nmake install

Running the Regression Tests

To verify a correct installation, type

make test

Use nmake test under Windows. There are a few demonstration scripts available to exercise many of the functions PerlMagick can perform. Type

Use the prove utility to execute a test from the build folder:

prove --blib blib -I `pwd` -bv ./t/read.t
cd demo
make

You are now ready to utilize the PerlMagick methods from within your Perl scripts.

Overview

Any script that wants to use PerlMagick methods must first define the methods within its namespace and instantiate an image object. Do this with:

use Image::Magick;

$image = Image::Magick->new;

PerlMagick is quantum aware. You can request a specific quantum depth when you instantiate an image object:

use Image::Magick::Q16;

$image = Image::Magick::Q16->new;

The new() method takes the same parameters as SetAttribute . For example,

$image = Image::Magick->new(size=>'384x256');

Next you will want to read an image or image sequence, manipulate it, and then display or write it. The input and output methods for PerlMagick are defined in Read or Write an Image. See Set an Image Attribute for methods that affect the way an image is read or written. Refer to Manipulate an Image for a list of methods to transform an image. Get an Image Attribute describes how to retrieve an attribute for an image. Refer to Create an Image Montage for details about tiling your images as thumbnails on a background. Finally, some methods do not neatly fit into any of the categories just mentioned. Review Miscellaneous Methods for a list of these methods.

Once you are finished with a PerlMagick object you should consider destroying it. Each image in an image sequence is stored in virtual memory. This can potentially add up to mebibytes of memory. Upon destroying a PerlMagick object, the memory is returned for use by other Perl methods. The recommended way to destroy an object is with undef:

undef $image;

To delete all the images but retain the Image::Magick object use

@$image = ();

and finally, to delete a single image from a multi-image sequence, use

undef $image->[$x];

The next section illustrates how to use various PerlMagick methods to manipulate an image sequence.

Some of the PerlMagick methods require external programs such as Ghostscript. This may require an explicit path in your PATH environment variable to work properly. For example (in Linux),

$ENV{PATH}' . "='/../bin:/usr/bin:/usr/local/bin';

Example Script

Here is an example script to get you started:

#!/usr/local/bin/perl
use Image::Magick;
my($image, $x);
$image = Image::Magick->new; $x = $image->Read('girl.png', 'logo.png', 'rose.png'); warn "$x" if "$x";
$x = $image->Crop(geometry=>'100x100+100+100'); warn "$x" if "$x";
$x = $image->Write('x.png'); warn "$x" if "$x";

The script reads three images, crops them, and writes a single image as a GIF animation sequence. In many cases you may want to access individual images of a sequence. The next example illustrates how this done:

#!/usr/local/bin/perl
use Image::Magick;
my($image, $p, $q);
$image = new Image::Magick; $image->Read('x1.png'); $image->Read('j*.jpg'); $image->Read('k.miff[1, 5, 3]'); $image->Contrast(); for ($x = 0; $image->[$x]; $x++) { $image->[$x]->Frame('100x200') if $image->[$x]->Get('magick') eq 'GIF'; undef $image->[$x] if $image->[$x]->Get('columns') < 100; } $p = $image->[1]; $p->Draw(stroke=>'red', primitive=>'rectangle', points=>20,20 100,100'); $q = $p->Montage(); undef $image; $q->Write('x.miff');

Suppose you want to start out with a 100 by 100 pixel white canvas with a red pixel in the center. Try

$image = Image::Magick->new;
$image->Set(size=>'100x100');
$image->ReadImage('canvas:white');
$image->Set('pixel[49,49]'=>'red');

Here we reduce the intensity of the red component at (1,1) by half:

@pixels = $image->GetPixel(x=>1,y=>1);
$pixels[0]*=0.5;
$image->SetPixel(x=>1,y=>1,color=>\@pixels);

Or suppose you want to convert your color image to grayscale:

$image->Quantize(colorspace=>'gray');

Let's annotate an image with a Taipai TrueType font:

$text = 'Works like magick!';
$image->Annotate(font=>'kai.ttf', pointsize=>40, fill=>'green', text=>$text);

Perhaps you want to extract all the pixel intensities from an image and write them to STDOUT:

@pixels = $image->GetPixels(map=>'I', height=>$height, width=>$width, normalize=>true);
binmode STDOUT;
print pack('B*',join('',@pixels));

Other clever things you can do with a PerlMagick objects include

$i = $#$p"+1";   # return the number of images associated with object p
push(@$q, @$p);  # push the images from object p onto object q
@$p = ();        # delete the images but not the object p
$p->Convolve([1, 2, 1, 2, 4, 2, 1, 2, 1]);   # 3x3 Gaussian kernel

Read or Write an Image

Use the methods listed below to either read, write, or display an image or image sequence:

Read or Write Methods
Method Parameters Return Value Description
Read one or more filenames the number of images read read an image or image sequence
Write filename the number of images written write an image or image sequence
Display server name the number of images displayed display the image or image sequence to an X server
Animate server name the number of images animated animate image sequence to an X server

For convenience, methods Write(), Display(), and Animate() can take any parameter that SetAttribute knows about. For example,

$image->Write(filename=>'image.png', compression=>'None');

Use - as the filename to method Read() to read from standard in or to method Write() to write to standard out:

binmode STDOUT;
$image->Write('png:-');

To read an image in the GIF format from a PERL filehandle, use:

$image = Image::Magick->new;
open(IMAGE, 'image.gif');
$image->Read(file=>\*IMAGE);
close(IMAGE);

To write an image in the PNG format to a PERL filehandle, use:

$filename = "image.png";
open(IMAGE, ">$filename");
$image->Write(file=>\*IMAGE, filename=>$filename);
close(IMAGE);

Note, reading from or writing to a Perl filehandle may fail under Windows due to different versions of the C-runtime libraries between ImageMagick and the ActiveState Perl distributions or if one of the DLL's is linked with the /MT option. See Potential Errors Passing CRT Objects Across DLL Boundaries for an explanation.

If %0Nd, %0No, or %0Nx appears in the filename, it is interpreted as a printf format specification and the specification is replaced with the specified decimal, octal, or hexadecimal encoding of the scene number. For example,

image%03d.miff

converts files image000.miff, image001.miff, etc.

You can optionally add Image to any method name. For example, ReadImage() is an alias for method Read().

Manipulate an Image

Once you create an image with, for example, method ReadImage() you may want to operate on it. Below is a list of all the image manipulations methods available to you with PerlMagick. There are examples of select PerlMagick methods. Here is an example call to an image manipulation method:

$image->Crop(geometry=>'100x100+10+20');
$image->[$x]->Frame("100x200");

And here is a list of other image manipulation methods you can call:

Image Manipulation Methods
Method Parameters Description
AdaptiveBlur geometry=>geometry, radius=>double, sigma=>double, bias=>double, channel=>{All, Default, Alpha, Black, Blue, CMYK, Cyan, Gray, Green, Index, Magenta, Alpha, Red, RGB, Yellow} adaptively blur the image with a Gaussian operator of the given radius and standard deviation (sigma). Decrease the effect near edges.
AdaptiveResize geometry=>geometry, width=>integer, height=>integer, filter=>{Point, Box, Triangle, Hermite, Hanning, Hamming, Blackman, Gaussian, Quadratic, Cubic, Catrom, Mitchell, Lanczos, Bessel, Sinc}, support=>double, blur=>double adaptively resize image using data dependant triangulation. Specify blur > 1 for blurry or < 1 for sharp
AdaptiveSharpen geometry=>geometry, radius=>double, sigma=>double, bias=>double, channel=>{All, Default, Alpha, Black, Blue, CMYK, Cyan, Gray, Green, Index, Magenta, Alpha, Red, RGB, Yellow} adaptively sharpen the image with a Gaussian operator of the given radius and standard deviation (sigma). Increase the effect near edges.
AdaptiveThreshold geometry=>geometry, width=>integer, height=>integer, bias=>double local adaptive thresholding.
AddNoise noise=>{Uniform, Gaussian, Multiplicative, Impulse, Laplacian, Poisson}, attenuate=>double, channel=>{All, Default, Alpha, Black, Blue, CMYK, Cyan, Gray, Green, Index, Magenta, Alpha, Red, RGB, Yellow} add noise to an image
AffineTransform affine=>array of float values, translate=>float, float, scale=> float, float, rotate=>float, skewX=>float, skewY=>float, interpolate={Average, Bicubic, Bilinear, Filter, Integer, Mesh, NearestNeighbor}, background=>color name affine transform image
Affinity image=>image-handle, method=>{None, FloydSteinberg, Riemersma} choose a particular set of colors from this image
Annotate text=>string, font=>string, family=>string, style=>{Normal, Italic, Oblique, Any}, stretch=>{Normal, UltraCondensed, ExtraCondensed, Condensed, SemiCondensed, SemiExpanded, Expanded, ExtraExpanded, UltraExpanded}, weight=>integer, pointsize=>integer, density=>geometry, stroke=>color name, strokewidth=>integer, fill=>color name, undercolor=>color name, kerning=>float, geometry=>geometry, gravity=>{NorthWest, North, NorthEast, West, Center, East, SouthWest, South, SouthEast}, antialias=>{true, false}, x=>integer, y=>integer, affine=>array of float values, translate=>float, float, scale=>float, float, rotate=>float. skewX=>float, skewY=> float, align=>{Left, Center, Right}, encoding=>{UTF-8}, interline-spacing=>double, interword-spacing=>double, direction=>{right-to-left, left-to-right}, decorate=>{none, underline, overline, line-through}, word-break=>{normal, break-word} annotate an image with text. See QueryFontMetrics to get font metrics without rendering any text.
AutoGamma channel=>{All, Default, Alpha, Black, Blue, CMYK, Cyan, Gray, Green, Index, Magenta, Alpha, Red, RGB, Yellow} automagically adjust gamma level of image
AutoLevel channel=>{All, Default, Alpha, Black, Blue, CMYK, Cyan, Gray, Green, Index, Magenta, Alpha, Red, RGB, Yellow} automagically adjust color levels of image
AutoOrient
adjusts an image so that its orientation is suitable for viewing (i.e. top-left orientation)
AutoThreshold method=>{Kapur, OTSU, Triangle} automatically perform image thresholding
BilateralSmoothing geometry=>geometry, width=>integer, height=>integer, intensity-sigma=>double, spatial-sigma=>double, channel=>{All, Default, Alpha, Black, Blue, CMYK, Cyan, Gray, Green, Index, Magenta, Alpha, Red, RGB, Yellow} a non-linear, edge-preserving, and noise-reducing smoothing filter for images. It replaces the intensity of each pixel with a weighted average of intensity values from nearby pixels. This weight is based on a Gaussian distribution. The weights depend not only on Euclidean distance of pixels, but also on the radiometric differences (e.g., range differences, such as color intensity, depth distance, etc.). This preserves sharp edges. The default value for the intensity and spatial sigmas are 2*diameter and 0.5*diameter respectively.
BlackThreshold threshold=>color, channel=>{All, Default, Alpha, Black, Blue, CMYK, Cyan, Gray, Green, Index, Magenta, Alpha, Red, RGB, Yellow} force all pixels below the threshold intensity into black
BlueShift factor=>double, simulate a scene at nighttime in the moonlight. Start with a factor of 1.5.
Blur geometry=>geometry, radius=>double, sigma=>double, bias=>double, channel=>{All, Default, Alpha, Black, Blue, CMYK, Cyan, Gray, Green, Index, Magenta, Alpha, Red, RGB, Yellow} reduce image noise and reduce detail levels with a Gaussian operator of the given radius and standard deviation (sigma).
Border geometry=>geometry, width=>integer, height=>integer, bordercolor=>color name, compose=>{Undefined, Add, Atop, Blend, Bumpmap, Clear, ColorBurn, ColorDodge, Colorize, CopyBlack, CopyBlue, CopyCMYK, Cyan, CopyGreen, Copy, CopyMagenta, CopyAlpha, CopyRed, RGB, CopyYellow, Darken, Dst, Difference, Displace, Dissolve, DstAtop, DstIn, DstOut, DstOver, Dst, Exclusion, HardLight, Hue, In, Lighten, Luminize, Minus, Modulate, Multiply, None, Out, Overlay, Over, Plus, ReplaceCompositeOp, Saturate, Screen, SoftLight, Src, SrcAtop, SrcIn, SrcOut, SrcOver, Src, Subtract, Threshold, Xor }, surround the image with a border of color
CannyEdge geometry=>geometry, radius=>double, sigma=>double, 'lower-percent'=>double, 'upper-percent'=>double use a multi-stage algorithm to detect a wide range of edges in the image (e.g. CannyEdge('0x1+10%+40%')).
Charcoal geometry=>geometry, radius=>double, sigma=>double simulate a charcoal drawing
Chop geometry=>geometry, width=>integer, height=>integer, x=>integer, y=>integer, gravity=>{NorthWest, North, NorthEast, West, Center, East, SouthWest, South, SouthEast} chop an image
CLAHE geometry=>geometry, width=>integer, height=>integer, number-bins=>integer, clip-limit=>double contrast limited adaptive histogram equalization. width, height divides the image into tiles. number-bins is the number of histogram bins per tile (min 2, max 256). clip-limit is the contrast limit for localised changes in contrast. A clip-limit of 2 to 3 is a good starting place.
Clamp channel=>{Red, RGB, All, etc.} set each pixel whose value is below zero to zero and any the pixel whose value is above the quantum range to the quantum range (e.g. 65535) otherwise the pixel value remains unchanged.
Clip id=>name, inside=>{true, false}, apply along a named path from the 8BIM profile.
ClipMask mask=>image-handle clip image as defined by the image mask
Clut image=>image-handle, interpolate={Average, Bicubic, Bilinear, Filter, Integer, Mesh, NearestNeighbor}, channel=>{Red, RGB, All, etc.} apply a color lookup table to an image sequence
Color color=>color name set the entire image to this color.
ColorDecisionList filename=>string, color correct with a color decision list.
Colorize fill=>color name, blend=>string colorize the image with the fill color
ColorMatrix matrix=>array of float values apply color correction to the image. Although you can use variable sized matrices, typically you use a 5 x 5 for an RGBA image and a 6x6 for CMYKA. A 6x6 matrix is required for offsets (populate the last column with normalized values).
Colorspace colorspace=>{RGB, Gray, Transparent, OHTA, XYZ, YCbCr, YCC, YIQ, YPbPr, YUV, CMYK} set the image colorspace
Comment string add a comment to your image
ColorThreshold start-color=>color, stop-color=>color, channel=>{All, Default, Alpha, Black, Blue, CMYK, Cyan, Gray, Green, Index, Magenta, Alpha, Red, RGB, Yellow} force all pixels below the threshold intensity into black
CompareLayers method=>{any, clear, overlay} compares each image with the next in a sequence and returns the minimum bounding region of any pixel differences it discovers. Images do not have to be the same size, though it is best that all the images are coalesced (images are all the same size, on a flattened canvas, so as to represent exactly how a specific frame should look).
Composite image=>image-handle, compose=>{Undefined, Add, Atop, Blend, Bumpmap, Clear, ColorBurn, ColorDodge, Colorize, CopyBlack, CopyBlue, CopyCMYK, Cyan, CopyGreen, Copy, CopyMagenta, CopyAlpha, CopyRed, RGB, CopyYellow, Darken, Dst, Difference, Displace, Dissolve, DstAtop, DstIn, DstOut, DstOver, Dst, Exclusion, HardLight, Hue, In, Lighten, Luminize, Minus, Modulate, Multiply, None, Out, Overlay, Over, Plus, ReplaceCompositeOp, Saturate, Screen, SoftLight, Src, SrcAtop, SrcIn, SrcOut, SrcOver, Src, Subtract, Threshold, Xor }, mask=>image-handle, geometry=>geometry, x=>integer, y=>integer, gravity=>{NorthWest, North, NorthEast, West, Center, East, SouthWest, South, SouthEast}, opacity=>integer, tile=>{True, False}, rotate=>double, color=>color name, blend=>geometry, interpolate=>{undefined, average, bicubic, bilinear, filter, integer, mesh, nearest-neighbor, spline}, clip-to-self=>{True, False} composite one image onto another. Use the rotate parameter in concert with the tile parameter.
ConnectedComponents connectivity=>integer, connected-components uniquely labeled, choose from 4 or 8 way connectivity.
Contrast sharpen=>{True, False} enhance or reduce the image contrast
ContrastStretch levels=>string, 'black-point'=>double, 'white-point'=>double, channel=>{Red, RGB, All, etc.} improve the contrast in an image by `stretching' the range of intensity values
Convolve coefficients=>array of float values, channel=>{All, Default, Alpha, Black, Blue, CMYK, Cyan, Gray, Green, Index, Magenta, Alpha, Red, RGB, Yellow}, bias=>double apply a convolution kernel to the image. Given a kernel order , you would supply order*order float values (e.g. 3x3 implies 9 values).
CopyPixels image=>image-handle, geometry=>geometry, width=>integer, height=>integer, x=>integer, y=>integer, offset=>geometry, gravity=>{NorthWest, North, NorthEast, West, Center, East, SouthWest, South, SouthEast}, dx=>integer, dy=>integer copy pixels from the image as defined by the widthxheight+x+y to image at offset +dx,+dy.
Crop geometry=>geometry, width=>integer, height=>integer, x=>integer, y=>integer, fuzz=>double, gravity=>{NorthWest, North, NorthEast, West, Center, East, SouthWest, South, SouthEast} crop an image
CycleColormap amount=>integer displace image colormap by amount
Decipher passphrase=>string convert cipher pixels to plain pixels
Deconstruct
break down an image sequence into constituent parts
Deskew geometry=>string,threshold=>double straighten the image
Despeckle reduce the speckles within an image
Difference image=>image-handle compute the difference metrics between two images
Distort points=>array of float values, method=>{Affine, AffineProjection, ScaleRotateTranslate, SRT, Perspective, PerspectiveProjection, BilinearForward, BilinearReverse, Polynomial, Arc, Polar, DePolar, Barrel, BarrelInverse, Shepards, Resize}, 'virtual-pixel'=>{Background Black Constant Dither Edge Gray Mirror Random Tile Transparent White}, 'best-fit'=>{True, False} distort image
Draw primitive=>{point, line, rectangle, arc, ellipse, circle, path, polyline, polygon, bezier, color, matte, text, @filename}, points=>string , method=>{Point, Replace, Floodfill, FillToBorder, Reset}, stroke=>color name, fill=>color name, font=>string, pointsize=>integer, strokewidth=>float, antialias=>{true, false}, bordercolor=>color name, x=>float, y=>float, dash-offset=>float, dash-pattern=>array of float values, affine=>array of float values, translate=>float, float, scale=>float, float, rotate=>float, skewX=>float, skewY=>float, interpolate=>{undefined, average, bicubic, bilinear, mesh, nearest-neighbor, spline}, kerning=>float, text=>string, vector-graphics=>string, interline-spacing=>double, interword-spacing=>double, direction=>{right-to-left, left-to-right}, word-break=>{normal, break-word} annotate an image with one or more graphic primitives.
Encipher passphrase=>string convert plain pixels to cipher pixels
Edge radius=>double enhance edges within the image with a convolution filter of the given radius.
Emboss geometry=>geometry, radius=>double, sigma=>double emboss the image with a convolution filter of the given radius and standard deviation (sigma).
Enhance
apply a digital filter to enhance a noisy image
Equalize channel=>{All, Default, Alpha, Black, Blue, CMYK, Cyan, Gray, Green, Index, Magenta, Alpha, Red, RGB, Yellow}
perform histogram equalization to the image
Extent geometry=>geometry, width=>integer, height=>integer, x=>integer, y=>integer, fuzz=>double, background=>color name, gravity=>{NorthWest, North, NorthEast, West, Center, East, SouthWest, South, SouthEast} set the image size
Evaluate value=>double, operator=>{Add, And, Divide, LeftShift, Max, Min, Multiply, Or, Rightshift, RMS, Subtract, Xor}, channel=>{All, Default, Alpha, Black, Blue, CMYK, Cyan, Gray, Green, Index, Magenta, Alpha, Red, RGB, Yellow} apply an arithmetic, relational, or logical expression to the image
Filter kernel=>string, channel=>{All, Default, Alpha, Black, Blue, CMYK, Cyan, Gray, Green, Index, Magenta, Alpha, Red, RGB, Yellow}, bias=>double apply a convolution kernel to the image.
Flip
reflect the image scanlines in the vertical direction
Flop
reflect the image scanlines in the horizontal direction
FloodfillPaint geometry=>geometry, channel=>{All, Default, Alpha, Black, Blue, CMYK, Cyan, Gray, Green, Index, Magenta, Alpha, Red, RGB, Yellow}, x=>integer, y=>integer , fill=>color name, bordercolor=>color name, fuzz=>double, invert=>{True, False} changes the color value of any pixel that matches the color of the target pixel and is a neighbor. If you specify a border color, the color value is changed for any neighbor pixel that is not that color.
ForwardFourierTransform magnitude=>{True, False} implements the forward discrete Fourier transform (DFT)
Frame geometry=>geometry, width=>integer, height=>integer, inner=>integer, outer=>integer, fill=>color name, compose=>{Undefined, Add, Atop, Blend, Bumpmap, Clear, ColorBurn, ColorDodge, Colorize, CopyBlack, CopyBlue, CopyCMYK, Cyan, CopyGreen, Copy, CopyMagenta, CopyAlpha, CopyRed, RGB, CopyYellow, Darken, Dst, Difference, Displace, Dissolve, DstAtop, DstIn, DstOut, DstOver, Dst, Exclusion, HardLight, Hue, In, Lighten, Luminize, Minus, Modulate, Multiply, None, Out, Overlay, Over, Plus, ReplaceCompositeOp, Saturate, Screen, SoftLight, Src, SrcAtop, SrcIn, SrcOut, SrcOver, Src, Subtract, Threshold, Xor }, surround the image with an ornamental border
Function parameters=>array of float values, function=>{Sin}, 'virtual-pixel'=>{Background Black Constant Dither Edge Gray Mirror Random Tile Transparent White} apply a function to the image
Gamma gamma=>string, channel=>{All, Default, Alpha, Black, Blue, CMYK, Cyan, Gray, Green, Index, Magenta, Alpha, Red, RGB, Yellow} gamma correct the image
GaussianBlur geometry=>geometry, radius=>double, sigma=>double, bias=>double, channel=>{All, Default, Alpha, Black, Blue, CMYK, Cyan, Gray, Green, Index, Magenta, Alpha, Red, RGB, Yellow} reduce image noise and reduce detail levels with a Gaussian operator of the given radius and standard deviation (sigma).
GetPixel geometry=>geometry, channel=>{All, Default, Alpha, Black, Blue, CMYK, Cyan, Gray, Green, Index, Magenta, Alpha, Red, RGB, Yellow}, normalize=>{true, false}, x=>integer, y=>integer get a single pixel. By default normalized pixel values are returned.
GetPixels geometry=>geometry, width=>integer, height=>integer, x=>integer, y=>integer, map=>string, normalize=>{true, false} get image pixels as defined by the map (e.g. "RGB", "RGBA", etc.). By default non-normalized pixel values are returned.
Grayscale channel=>{Average, Brightness, Lightness, Rec601Luma, Rec601Luminance, Rec709Luma, Rec709Luminance, RMS} convert image to grayscale
HaldClut image=>image-handle, channel=>{Red, RGB, All, etc.} apply a Hald color lookup table to an image sequence
HoughLine geometry=>geometry, width=>double, height=>double, threshold=>double identify lines in the image (e.g. HoughLine('9x9+195')).
Identify file=>file, features=>distance, moments=>{True, False}, unique=>{True, False} identify the attributes of an image
Implode amount=>double, interpolate=>{undefined, average, bicubic, bilinear, mesh, nearest-neighbor, spline} implode image pixels about the center
Integral calculate the sum of values (pixel values) in the image.
InverseDiscreteFourierTransform magnitude=>{True, False} implements the inverse discrete Fourier transform (DFT)
Kmeans geometry=>geometry, 'colors'=>double, 'iterations'=>double, 'tolerance'=>double K means color reduction.
Kuwahara geometry=>geometry, radius=>double, sigma=>double, bias=>double, channel=>{All, Default, Alpha, Black, Blue, CMYK, Cyan, Gray, Green, Index, Magenta, Alpha, Red, RGB, Yellow} edge preserving noise reduction filter
Label string assign a label to an image
Layers method=>{coalesce, compare-any, compare-clear, compare-over, composite, dispose, flatten, merge, mosaic, optimize, optimize-image, optimize-plus, optimize-trans, remove-dups, remove-zero}, compose=>{Undefined, Add, Atop, Blend, Bumpmap, Clear, ColorBurn, ColorDodge, Colorize, CopyBlack, CopyBlue, CopyCMYK, Cyan, CopyGreen, Copy, CopyMagenta, CopyAlpha, CopyRed, RGB, CopyYellow, Darken, Dst, Difference, Displace, Dissolve, DstAtop, DstIn, DstOut, DstOver, Dst, Exclusion, HardLight, Hue, In, Lighten, LinearLight, Luminize, Minus, Modulate, Multiply, None, Out, Overlay, Over, Plus, ReplaceCompositeOp, Saturate, Screen, SoftLight, Src, SrcAtop, SrcIn, SrcOut, SrcOver, Src, Subtract, Threshold, Xor }, dither=>{true, false} compare each image the GIF disposed forms of the previous image in the sequence. From this, attempt to select the smallest cropped image to replace each frame, while preserving the results of the animation.
Level levels=>string, 'black-point'=>double, 'gamma'=>double, 'white-point'=>double, channel=>{Red, RGB, All, etc.} adjust the level of image contrast
LevelColors invert=>>{True, False}, 'black-point'=>string, 'white-point'=>string, channel=>{Red, RGB, All, etc.} level image with the given colors
LinearStretch levels=>string, 'black-point'=>double, 'white-point'=>double linear with saturation stretch
LiquidResize geometry=>geometry, width=>integer, height=>integer, delta-x=>double, rigidity=>double rescale image with seam-carving.
Magnify
double the size of the image with pixel art scaling
Mask mask=>image-handle composite image pixels as defined by the mask
MatteFloodfill geometry=>geometry, x=>integer, y=>integer , matte=>integer, bordercolor=>color name, fuzz=>double, invert=>{True, False} changes the matte value of any pixel that matches the color of the target pixel and is a neighbor. If you specify a border color, the matte value is changed for any neighbor pixel that is not that color.
MeanShift geometry=>geometry, width=>double, height=>double, distance=>double delineate arbitrarily shaped clusters in the image (e.g. MeanShift('7x7+10%')).
MedianFilter geometry=>geometry, width=>integer, height=>integer, channel=>{All, Default, Alpha, Black, Blue, CMYK, Cyan, Gray, Green, Index, Magenta, Alpha, Red, RGB, Yellow} replace each pixel with the median intensity pixel of a neighborhood.
Minify
half the size of an image
Mode geometry=>geometry, width=>integer, height=>integer, channel=>{All, Default, Alpha, Black, Blue, CMYK, Cyan, Gray, Green, Index, Magenta, Alpha, Red, RGB, Yellow} make each pixel the predominant color of the neighborhood.
Modulate factor=>geometry, brightness=>double, saturation=>double, hue=>double, lightness=>double, whiteness=>double, blackness=>double vary the brightness, saturation, and hue of an image by the specified percentage
Morphology kernel=>string, channel=>{All, Default, Alpha, Black, Blue, CMYK, Cyan, Gray, Green, Index, Magenta, Alpha, Red, RGB, Yellow}, iterations=>integer apply a morphology method to the image.
MotionBlur geometry=>geometry, radius=>double, sigma=>double, angle=>double, bias=>double, channel=>{All, Default, Alpha, Black, Blue, CMYK, Cyan, Gray, Green, Index, Magenta, Alpha, Red, RGB, Yellow} reduce image noise and reduce detail levels with a Gaussian operator of the given radius and standard deviation (sigma) at the given angle to simulate the effect of motion
Negate gray=>{True, False}, channel=>{All, Default, Alpha, Black, Blue, CMYK, Cyan, Gray, Green, Index, Magenta, Alpha, Red, RGB, Yellow} replace each pixel with its complementary color (white becomes black, yellow becomes blue, etc.)
Normalize channel=>{All, Default, Alpha, Black, Blue, CMYK, Cyan, Gray, Green, Index, Magenta, Alpha, Red, RGB, Yellow}
transform image to span the full range of color values
OilPaint radius=>integer simulate an oil painting
Opaque color=>color name, fill=>color name, channel=>{All, Default, Alpha, Black, Blue, CMYK, Cyan, Gray, Green, Index, Magenta, Alpha, Red, RGB, Yellow}, invert=>{True, False} change this color to the fill color within the image
OrderedDither threshold=>{threshold, checks, o2x2, o3x3, o4x4, o8x8, h4x4a, h6x6a, h8x8a, h4x4o, h6x6o, h8x8o, h16x16o, hlines6x4}, channel=>{All, Default, Alpha, Black, Blue, CMYK, Cyan, Gray, Green, Index, Magenta, Alpha, Red, RGB, Yellow} order dither image
Perceptible epsilon=>double, channel=>{Red, RGB, All, etc.} set each pixel whose value is less than |epsilon| to -epsilon or epsilon (whichever is closer) otherwise the pixel value remains unchanged..
Polaroid caption=>string, angle=>double, pointsize=>double, font=>string, stroke=> color name, strokewidth=>integer, fill=>color name, gravity=>{NorthWest, North, NorthEast, West, Center, East, SouthWest, South, SouthEast}, background=>color name simulate a Polaroid picture.
Posterize levels=>integer, dither=>{True, False} reduce the image to a limited number of color level
Profile name=>string, profile=>blob, rendering-intent=>{Undefined, Saturation, Perceptual, Absolute, Relative}, black-point-compensation=>{True, False} add or remove ICC or IPTC image profile; name is formal name (e.g. ICC or filename; set profile to '' to remove profile
Quantize colors=>integer, colorspace=>{RGB, Gray, Transparent, OHTA, XYZ, YCbCr, YIQ, YPbPr, YUV, CMYK, sRGB, HSL, HSB}, treedepth=> integer, dither=>{True, False}, dither-method=>{Riemersma, Floyd-Steinberg}, measure_error=>{True, False}, global_colormap=>{True, False}, transparent-color=>color preferred number of colors in the image
Raise geometry=>geometry, width=>integer, height=>integer, x=>integer, y=>integer, raise=>{True, False} lighten or darken image edges to create a 3-D effect
RangeThreshold geometry=>geometry, 'low-black'=>double, 'low-white'=>double, 'high-white'=>double, 'high-black'=>double combine soft and hard image thresholding.
ReduceNoise geometry=>geometry, width=>integer, height=>integer, channel=>{All, Default, Alpha, Black, Blue, CMYK, Cyan, Gray, Green, Index, Magenta, Alpha, Red, RGB, Yellow} reduce noise in the image with a noise peak elimination filter
Remap image=>image-handle, dither=>{true, false}, dither-method=>{Riemersma, Floyd-Steinberg} replace the colors of an image with the closest color from a reference image.
Resample density=>geometry, x=>double, y=>double, filter=>{Point, Box, Triangle, Hermite, Hanning, Hamming, Blackman, Gaussian, Quadratic, Cubic, Catrom, Mitchell, Lanczos, Bessel, Sinc}, support=>double resample image to desired resolution. Specify blur > 1 for blurry or < 1 for sharp
Resize geometry=>geometry, width=>integer, height=>integer, filter=>{Point, Box, Triangle, Hermite, Hanning, Hamming, Blackman, Gaussian, Quadratic, Cubic, Catrom, Mitchell, Lanczos, Bessel, Sinc}, support=>double, blur=>double scale image to desired size. Specify blur > 1 for blurry or < 1 for sharp
Roll geometry=>geometry, x=>integer, y=>integer roll an image vertically or horizontally
Rotate degrees=>double, background=>color name rotate an image
RotationalBlur geometry=>geometry, angle=>double, bias=>double, channel=>{All, Default, Alpha, Black, Blue, CMYK, Cyan, Gray, Green, Index, Magenta, Alpha, Red, RGB, Yellow} radial blur the image.
Sample geometry=>geometry, width=>integer, height=>integer scale image with pixel sampling.
Scale geometry=>geometry, width=>integer, height=>integer scale image to desired size
Segment colorspace=>{RGB, Gray, Transparent, OHTA, XYZ, YCbCr, YCC, YIQ, YPbPr, YUV, CMYK}, verbose={True, False}, cluster-threshold=>double, smoothing-threshold=double segment an image by analyzing the histograms of the color components and identifying units that are homogeneous
SelectiveBlur geometry=>geometry, radius=>double, sigma=>double, threshold=>double, bias=>double, channel=>{All, Default, Alpha, Black, Blue, CMYK, Cyan, Gray, Green, Index, Magenta, Alpha, Red, RGB, Yellow} selectively blur pixels within a contrast threshold.
Separate channel=>{Red, RGB, All, etc.} separate a channel from the image into a grayscale image
Shade geometry=>geometry, azimuth=>double, elevation=>double, gray=>{true, false} shade the image using a distant light source
SetPixel geometry=>geometry, channel=>{All, Default, Alpha, Black, Blue, CMYK, Cyan, Gray, Green, Index, Magenta, Alpha, Red, RGB, Yellow}, color=>array of float values, x=>integer, y=>integer, color=>array of float values set the value a single pixel. Normalized pixel values are expected.
SetPixels geometry=>geometry, channel=>{All, Default, Alpha, Black, Blue, CMYK, Cyan, Gray, Green, Index, Magenta, Alpha, Red, RGB, Yellow}, color=>array of float values, width=>integer, height=>integer, x=>integer, y=>integer, color=>array of float values set the value of one or more pixels. Normalized pixel values are expected.
Shadow geometry=>geometry, opacity=>double, sigma=>double, x=>integer, y=>integer simulate an image shadow
Sharpen geometry=>geometry, radius=>double, sigma=>double, bias=>double, channel=>{All, Default, Alpha, Black, Blue, CMYK, Cyan, Gray, Green, Index, Magenta, Alpha, Red, RGB, Yellow} sharpen the image with a Gaussian operator of the given radius and standard deviation (sigma).
Shave geometry=>geometry, width=>integer, height=>integer shave pixels from the image edges
Shear geometry=>geometry, x=>double, y=>double fill=>color name shear the image along the X or Y axis by a positive or negative shear angle
SigmoidalContrast geometry=>string, 'contrast'=>double, 'mid-point'=>double channel=>{Red, RGB, All, etc.}, sharpen=>{True, False} sigmoidal non-lineraity contrast control. Increase the contrast of the image using a sigmoidal transfer function without saturating highlights or shadows. Contrast indicates how much to increase the contrast (0 is none; 3 is typical; 20 is a lot); mid-point indicates where midtones fall in the resultant image (0 is white; 50% is middle-gray; 100% is black). To decrease contrast, set sharpen to False.
Signature
generate an SHA-256 message digest for the image pixel stream
Sketch geometry=>geometry, radius=>double, sigma=>double, angle=>double sketch the image with a Gaussian operator of the given radius and standard deviation (sigma) at the given angle
Solarize geometry=>string, threshold=>double, channel=>{All, Default, Alpha, Black, Blue, CMYK, Cyan, Gray, Green, Index, Magenta, Alpha, Red, RGB, Yellow} negate all pixels above the threshold level
SortPixels
sorts pixels within each scanline in ascending order of intensity.
SparseColor points=>array of float values, method=>{Barycentric, Bilinear, Shepards, Voronoi}, 'virtual-pixel'=>{Background Black Constant Dither Edge Gray Mirror Random Tile Transparent White} interpolate the image colors around the supplied points
Splice geometry=>geometry, width=>integer, height=>integer, x=>integer, y=>integer, fuzz=>double, background=>color name, gravity=>{NorthWest, North, NorthEast, West, Center, East, SouthWest, South, SouthEast} splice an image
Spread radius=>double, interpolate=>{undefined, average, bicubic, bilinear, mesh, nearest-neighbor, spline} displace image pixels by a random amount
Statistic geometry=>geometry, width=>integer, height=>integer, channel=>{All, Default, Alpha, Black, Blue, CMYK, Cyan, Gray, Green, Index, Magenta, Alpha, Red, RGB, Yellow}, type=>{Contrast, Median, Mode, Mean, Maximum, Minimum, ReduceNoise, RMS} replace each pixel with corresponding statistic from the neighborhood.
Stegano image=>image-handle, offset=>integer hide a digital watermark within the image
Stereo image=>image-handle, x=>integer, y=>integer composites two images and produces a single image that is the composite of a left and right image of a stereo pair
Strip
strip an image of all profiles and comments.
Swirl degrees=>double, interpolate=>{undefined, average, bicubic, bilinear, mesh, nearest-neighbor, spline} swirl image pixels about the center
Texture texture=>image-handle name of texture to tile onto the image background
Thumbnail geometry=>geometry, width=>integer, height=>integer changes the size of an image to the given dimensions and removes any associated profiles.
Threshold threshold=>string, channel=>{All, Default, Alpha, Black, Blue, CMYK, Cyan, Gray, Green, Index, Magenta, Alpha, Red, RGB, Yellow} threshold the image
Tint fill=>color name, blend=>string tint the image with the fill color.
Transparent color=>color name, invert=>{True, False} make this color transparent within the image
Transpose
flip image in the vertical direction and rotate 90 degrees
Transverse
flop image in the horizontal direction and rotate 270 degrees
Trim
remove edges that are the background color from the image
UnsharpMask geometry=>geometry, radius=>double, sigma=>double, gain=>double, threshold=>double sharpen the image with the unsharp mask algorithm.
Vignette geometry=>geometry, radius=>double, sigma=>double, x=>integer, y=>integer, background=>color name offset the edges of the image in vignette style
Wave geometry=>geometry, amplitude=>double, wavelength=>double, interpolate=>{undefined, average, bicubic, bilinear, mesh, nearest-neighbor, spline} alter an image along a sine wave
WaveDenoise geometry=>geometry, threshold=>double, threshold=>double removes noise from the image using a wavelet transform
WhiteBalance applies white balancing to an image according to a grayworld assumption in the LAB colorspace.
WhiteThreshold threshold=>string, , channel=>{All, Default, Alpha, Black, Blue, CMYK, Cyan, Gray, Green, Index, Magenta, Alpha, Red, RGB, Yellow} force all pixels above the threshold intensity into white

Note, that the geometry parameter is a short cut for the width and height parameters (e.g. geometry=>'106x80' is equivalent to width=>106, height=>80 ).

You can specify @filename in both Annotate() and Draw(). This reads the text or graphic primitive instructions from a file on disk. For example,

image->Draw(fill=>'red', primitive=>'rectangle',
 points=>'20,20 100,100  40,40 200,200  60,60 300,300');

Is equivalent to

$image->Draw(fill=>'red', primitive=>'@draw.txt');

Where draw.txt is a file on disk that contains this:

rectangle 20, 20 100, 100
rectangle 40, 40 200, 200
rectangle 60, 60 300, 300

The text parameter for methods, Annotate(), Comment(), Draw(), and Label() can include the image filename, type, width, height, or other image attribute by embedding these special format characters:

%b   file size
%c   comment
%d   directory
%e   filename extension
%f   filename
%g   page geometry
%h   height
%i   input filename
%k   number of unique colors
%l   label
%m   magick
%n   number of scenes
%o   output filename
%p   page number
%q   quantum depth
%r   image class and colorspace
%s   scene number
%t   top of filename
%u   unique temporary filename
%w   width
%x   x resolution
%y   y resolution
%z   image depth
%C   image compression type
%D   image dispose method
%H   page height
%Q   image compression quality
%T   image delay
%W   page width
%X   page x offset
%Y   page y offset
%@   bounding box
%#   signature
%%   a percent sign
\n   newline
\r   carriage return

For example,

text=>"%m:%f %wx%h"

produces an annotation of MIFF:bird.miff 512x480 for an image titled bird.miff and whose width is 512 and height is 480.

You can optionally add Image to any method name. For example, TrimImage() is an alias for method Trim().

Most of the attributes listed above have an analog in magick. See the documentation for a more detailed description of these attributes.

Set an Image Attribute

Use method Set() to set an image attribute. For example,

$image->Set(dither=>'True');
$image->[$x]->Set(delay=>3);

Where this example uses 'True' and this document says '{True, False}', you can use the case-insensitive strings 'True' and 'False', or you can use the integers 1 and 0.

When you call Get() on a Boolean attribute, Image::Magick returns 1 or 0, not a string.

And here is a list of all the image attributes you can set:

Image Attributes
Attribute Values Description
adjoin {True, False} join images into a single multi-image file
alpha {On, Off, Opaque, Transparent, Copy, Extract, Set} control of and special operations involving the alpha/matte channel
antialias {True, False} remove pixel aliasing
area-limit integer set pixel area resource limit.
attenuate double lessen (or intensify) when adding noise to an image.
authenticate string decrypt image with this password.
background color name image background color
blue-primary x-value, y-value chromaticity blue primary point (e.g. 0.15, 0.06)
bordercolor color name set the image border color
clip-mask image associate a clip mask with the image.
colormap[i] color name color name (e.g. red) or hex value (e.g. #ccc) at position i
comment string set the image comment
compression {None, BZip, Fax, Group4, JPEG, JPEG2000, LosslessJPEG, LZW, RLE, Zip} type of image compression
debug {All, Annotate, Blob, Cache, Coder, Configure, Deprecate, Draw, Exception, Locale, None, Resource, Transform, X11} display copious debugging information
delay integer this many 1/100ths of a second must expire before displaying the next image in a sequence
density geometry vertical and horizontal resolution in pixels of the image
depth integer image depth
direction {Undefined, right-to-left, left-to-right render text right-to-left or left-to-right
disk-limit integer set disk resource limit
dispose {Undefined, None, Background, Previous} layer disposal method
dither {True, False} apply error diffusion to the image
display string specifies the X server to contact
extract geometry extract area from image
file filehandle set the image filehandle
filename string set the image filename
fill color The fill color paints any areas inside the outline of drawn shape.
font string use this font when annotating the image with text
fuzz integer colors within this distance are considered equal
gamma double gamma level of the image
Gravity {Forget, NorthWest, North, NorthEast, West, Center, East, SouthWest, South, SouthEast} type of image gravity
green-primary x-value, y-value chromaticity green primary point (e.g. 0.3, 0.6)
index[x, y] string colormap index at position (x, y)
interlace {None, Line, Plane, Partition, JPEG, GIF, PNG} the type of interlacing scheme
iterations integer add Netscape loop extension to your GIF animation
label string set the image label
loop integer add Netscape loop extension to your GIF animation
magick string set the image format
map-limit integer set map resource limit
mask image associate a mask with the image.
matte {True, False} enable the image matte channel
mattecolor color name set the image matte color
memory-limit integer set memory resource limit
monochrome {True, False} transform the image to black and white
option string associate an option with an image format (e.g. option=>'ps:imagemask'
orientation {TopLeft, TopRight, BottomRight, BottomLeft, LeftTop, RightTop, RightBottom, LeftBottom} image orientation
page { Letter, Tabloid, Ledger, Legal, Statement, Executive, A3, A4, A5, B4, B5, Folio, Quarto, 10x14} or geometry preferred size and location of an image canvas
pixel[x, y] string hex value (e.g. #ccc) at position (x, y)
pointsize integer pointsize of the Postscript or TrueType font
precision integer set the maximum number of significant digits to be printed
quality integer JPEG/MIFF/PNG compression level
red-primary x-value, y-value chromaticity red primary point (e.g. 0.64, 0.33)
sampling-factor geometry horizontal and vertical sampling factor
scene integer image scene number
server string specifies the X server to contact
size string width and height of a raw image
stroke color The stroke color paints along the outline of a shape.
texture string name of texture to tile onto the image background
tile-offset geometry image tile offset
time-limit integer set time resource limit in seconds
title string set the image title
type {Bilevel, Grayscale, GrayscaleMatte, Palette, PaletteMatte, TrueColor, TrueColorMatte, ColorSeparation, ColorSeparationMatte} image type
units { Undefined, PixelsPerInch, PixelsPerCentimeter} units of image resolution
verbose {True, False} print detailed information about the image
virtual-pixel {Background Black Constant Dither Edge Gray Mirror Random Tile Transparent White} the virtual pixel method
white-point x-value, y-value chromaticity white point (e.g. 0.3127, 0.329)

Note, that the geometry parameter is a short cut for the width and height parameters (e.g. geometry=>'106x80' is equivalent to width=>106, height=>80).

SetAttribute() is an alias for method Set().

Most of the attributes listed above have an analog in magick. See the documentation for a more detailed description of these attributes.

Get an Image Attribute

Use method Get() to get an image attribute. For example,

($a, $b, $c) = $image->Get('colorspace', 'magick', 'adjoin');
$width = $image->[3]->Get('columns');

In addition to all the attributes listed in Set an Image Attribute , you can get these additional attributes:

Image Attributes
Attribute Values Description
area integer current area resource consumed
base-columns integer base image width (before transformations)
base-filename string base image filename (before transformations)
base-rows integer base image height (before transformations)
class {Direct, Pseudo} image class
colors integer number of unique colors in the image
columns integer image width
copyright string get PerlMagick's copyright
directory string tile names from within an image montage
elapsed-time double elapsed time in seconds since the image was created
error double the mean error per pixel computed with methods Compare() or Quantize()
bounding-box string image bounding box
disk integer current disk resource consumed
filesize integer number of bytes of the image on disk
format string get the descriptive image format
geometry string image geometry
height integer the number of rows or height of an image
icc string ICC profile
icc string ICM profile
id integer ImageMagick registry id
IPTC string IPTC profile
mean-error double the normalized mean error per pixel computed with methods Compare() or Quantize()
map integer current memory-mapped resource consumed
matte {True, False} whether or not the image has a matte channel
maximum-error double the normalized max error per pixel computed with methods Compare() or Quantize()
memory integer current memory resource consumed
mime string MIME of the image format
montage geometry tile size and offset within an image montage
page.x integer x offset of image virtual canvas
page.y integer y offset of image virtual canvas
rows integer the number of rows or height of an image
signature string SHA-256 message digest associated with the image pixel stream
taint {True, False} True if the image has been modified
total-ink-density double returns the total ink density for a CMYK image
transparent-color color name set the image transparent color
user-time double user time in seconds since the image was created
version string get PerlMagick's version
width integer the number of columns or width of an image
XMP string XMP profile
x-resolution integer x resolution of the image
y-resolution integer y resolution of the image

GetAttribute() is an alias for method Get().

Most of the attributes listed above have an analog in magick. See the documentation for a more detailed description of these attributes.

Compare an Image to its Reconstruction

Mathematically and visually annotate the difference between an image and its reconstruction with the Compare() method. The method supports these parameters:

Compare Parameters
Parameter Values Description
channel double select image channels, the default is all channels except alpha.
fuzz double colors within this distance are considered equal
image image-reference the image reconstruction
metric AE, MAE, MEPP, MSE, PAE, PSNR, RMSE measure differences between images with this metric

In this example, we compare the ImageMagick logo to a sharpened reconstruction:

use Image::Magick;

$logo=Image::Magick->New();
$logo->Read('logo:');
$sharp=Image::Magick->New();
$sharp->Read('logo:');
$sharp->Sharpen('0x1');
$difference=$logo->Compare(image=>$sharp, metric=>'rmse');
print $difference->Get('error'), "\n";
$difference->Display();

In addition to the reported root mean squared error of around 0.024, a difference image is displayed so you can visually identify the difference between the images.

Create an Image Montage

Use method Montage() to create a composite image by combining several separate images. The images are tiled on the composite image with the name of the image optionally appearing just below the individual tile. For example,

$image->Montage(geometry=>'160x160', tile=>'2x2', texture=>'granite:');

And here is a list of Montage() parameters you can set:

Montage Parameters
Parameter Values Description
background color name background color name
border integer image border width
filename string name of montage image
fill color name fill color for annotations
font string X11 font name
frame geometry surround the image with an ornamental border
geometry geometry preferred tile and border size of each tile of the composite image (e.g. 120x120+4+3>)
gravity NorthWest, North, NorthEast, West, Center, East, SouthWest, South, SouthEast direction image gravitates to within a tile
label string assign a label to an image
mode Frame, Unframe, Concatenate thumbnail framing options
pointsize integer pointsize of the Postscript or TrueType font
shadow {True, False} add a shadow beneath a tile to simulate depth
stroke color name stroke color for annotations
texture string name of texture to tile onto the image background
tile geometry the number of tiles per row and page (e.g. 6x4)
title string assign a title to the image montage
transparent string make this color transparent within the image

Note, that the geometry parameter is a short cut for the width and height parameters (e.g. geometry=>'106x80' is equivalent to width=>106, height=>80).

MontageImage() is an alias for method Montage().

Most of the attributes listed above have an analog in montage. See the documentation for a more detailed description of these attributes.

Working with Blobs

A blob contains data that directly represent a particular image format in memory instead of on disk. PerlMagick supports blobs in any of these image formats and provides methods to convert a blob to or from a particular image format.

Blob Methods
Method Parameters Return Value Description
ImageToBlob any image attribute an array of image data in the respective image format convert an image or image sequence to an array of blobs
BlobToImage one or more blobs the number of blobs converted to an image convert one or more blobs to an image

ImageToBlob() returns the image data in their respective formats. You can then print it, save it to an ODBC database, write it to a file, or pipe it to a display program:

@blobs = $image->ImageToBlob();
open(DISPLAY,"| display -") || die;
binmode DISPLAY;
print DISPLAY $blobs[0];
close DISPLAY;

Method BlobToImage() returns an image or image sequence converted from the supplied blob:

@blob=$db->GetImage();
$image=Image::Magick->new(magick=>'jpg');
$image->BlobToImage(@blob);

Direct-access to Image Pixels

Use these methods to obtain direct access to the image pixels:

Direct-access to Image Pixels
Method Parameters Description
GetAuthenticPixels geometry=>geometry, width=>integer, height=>integer, x=>integer, y=>integer return authentic pixels as a C pointer
GetVirtualPixels geometry=>geometry, width=>integer, height=>integer, x=>integer, y=>integer return virtual pixels as a const C pointer
GetAuthenticIndexQueue return colormap indexes or black pixels as a C pointer
GetVirtualIndexQueue return colormap indexes or black pixels as a const C pointer
SyncAuthenticPixels sync authentic pixels to pixel cache

Miscellaneous Methods

The Append() method append a set of images. For example,

$p = $image->Append(stack=>{true,false});

appends all the images associated with object $image. By default, images are stacked left-to-right. Set stack to True to stack them top-to-bottom.

The Clone() method copies a set of images. For example,

$q = $p->Clone();

copies all the images from object $p to $q. You can use this method for single or multi-image sequences.

Coalesce() composites a set of images while respecting any page offsets and disposal methods. GIF, MIFF, and MNG animation sequences typically start with an image background and each subsequent image varies in size and offset. A new image sequence is returned with all images the same size as the first images virtual canvas and composited with the next image in the sequence.. For example,

$q = $p->Coalesce();

The ComplexImages() method performs complex mathematics on an image sequence. For example,

$p = $image->ComplexImages('conjugate');

The EvaluateImages() method applies an arithmetic, logical or relational expression to a set of images. For example,

$p = $image->EvaluateImages('mean');

averages all the images associated with object $image.

The Features() method returns features for each channel in the image in each of four directions (horizontal, vertical, left and right diagonals) for the specified distance. The features include the angular second momentum, contrast, correlation, sum of squares: variance, inverse difference moment, sum average, sum varience, sum entropy, entropy, difference variance, difference entropy, information measures of correlation 1, information measures of correlation 2, and maximum correlation coefficient. Values in RGB, CMYK, RGBA, or CMYKA order (depending on the image type).

@features = $image->Features(1);

The Flatten() method flattens a set of images and returns it. For example,

$p = $images->Flatten(background=>'none');
$p->Write('flatten.png');

The sequence of images is replaced by a single image created by composing each image after the first over the first image.

The Fx() method applies a mathematical expression to a set of images and returns the results. For example,

$p = $image->Fx(expression=>'(g+b)/2.0',channel=>'red');
$p->Write('fx.miff');

replaces the red channel with the average of the green and blue channels.

See FX, The Special Effects Image Operator for a detailed discussion of this method.

Histogram() returns the unique colors in the image and a count for each one. The returned values are an array of red, green, blue, opacity, and count values.

The Morph() method morphs a set of images. Both the image pixels and size are linearly interpolated to give the appearance of a meta-morphosis from one image to the next:

$p = $image->Morph(frames=>integer);

where frames is the number of in-between images to generate. The default is 1.

Mosaic() creates an mosaic from an image sequence.

Method Mogrify() is a single entry point for the image manipulation methods (Manipulate an Image). The parameters are the name of a method followed by any parameters the method may require. For example, these calls are equivalent:

$image->Crop('340x256+0+0');
$image->Mogrify('crop', '340x256+0+0');

Method MogrifyRegion() applies a transform to a region of the image. It is similar to Mogrify() but begins with the region geometry. For example, suppose you want to brighten a 100x100 region of your image at location (40, 50):

$image->MogrifyRegion('100x100+40+50', 'modulate', brightness=>50);

PerceptualHash() maps visually identical images to the same or similar hash-- useful in image retrieval, authentication, indexing, or copy detection as well as digital watermarking. For each channel and for the sRGB and the HCLp colorspaces, 7 hash values are returned For an sRGB images, for example, expect 42 perceptual hashes.

@phash = $image->PerceptualHash();

Ping() is a convenience method that returns information about an image without having to read the image into memory. It returns the width, height, file size in bytes, and the file format of the image. You can specify more than one filename but only one filehandle:

($width, $height, $size, $format) = $image->Ping('logo.png');
($width, $height, $size, $format) = $image->Ping(file=>\*IMAGE);
($width, $height, $size, $format) = $image->Ping(blob=>$blob);

This a more efficient and less memory intensive way to query if an image exists and what its characteristics are.

Poly() builds a polynomial from the image sequence and the corresponding terms (coefficients and degree pairs):

$p = $image->Poly([0.5,1.0,0.25,2.0,1.0,1.0]);

PreviewImage() tiles 9 thumbnails of the specified image with an image processing operation applied at varying strengths. This may be helpful pin-pointing an appropriate parameter for a particular image processing operation. Choose from these operations: Rotate, Shear, Roll, Hue, Saturation, Brightness, Gamma, Spiff, Dull, Grayscale, Quantize, Despeckle, ReduceNoise, AddNoise, Sharpen, Blur, Threshold, EdgeDetect, Spread, Solarize, Shade, Raise, Segment, Swirl, Implode, Wave, OilPaint, CharcoalDrawing, JPEG. Here is an example:

$preview = $image->Preview('Gamma');
$preview->Display();

To have full control over text positioning you need font metric information. Use

($x_ppem, $y_ppem, $ascender, $descender, $width, $height, $max_advance) =
  $image->QueryFontMetrics(parameters);

Where parameters is any parameter of the Annotate method. The return values are:

  1. character width
  2. character height
  3. ascender
  4. descender
  5. text width
  6. text height
  7. maximum horizontal advance
  8. bounds: x1
  9. bounds: y1
  10. bounds: x2
  11. bounds: y2
  12. origin: x
  13. origin: y

Use QueryMultilineFontMetrics() to get the maximum text width and height for multiple lines of text.

Call QueryColor() with no parameters to return a list of known colors names or specify one or more color names to get these attributes: red, green, blue, and opacity value.

@colors = $image->QueryColor();
($red, $green, $blue) = $image->QueryColor('cyan');
($red, $green, $blue, $alpha) = $image->QueryColor('#716baeff');

QueryColorname() accepts a color value and returns its respective name or hex value;

$name = $image->QueryColorname('rgba(80,60,0,0)');

Call QueryFont() with no parameters to return a list of known fonts or specify one or more font names to get these attributes: font name, description, family, style, stretch, weight, encoding, foundry, format, metrics, and glyphs values.

@fonts = $image->QueryFont();
$weight = ($image->QueryFont('Helvetica'))[5];

Call QueryFormat() with no parameters to return a list of known image formats or specify one or more format names to get these attributes: adjoin, blob support, raw, decoder, encoder, description, and module.

@formats = $image->QueryFormat();
($adjoin, $blob_support, $raw, $decoder, $encoder, $description, $module) =
  $image->QueryFormat('gif');

Call MagickToMime() with the image format name to get its MIME type such as images/tiff from tif.

$mime = $image->MagickToMime('tif');

Use RemoteCommand() to send a command to an already running display or animate application. The only parameter is the name of the image file to display or animate.

$image->RemoteCommand('image.jpg');

The Smush() method smushes a set of images together. For example,

$p = $image->Smush(stack=>{true,false},offset=>integer);

smushes together all the images associated with object $image. By default, images are smushed left-to-right. Set stack to True to smushed them top-to-bottom.

Statistics() returns the image statistics for each channel in the image. The returned values are an array of depth, minima, maxima, mean, standard deviation, kurtosis, skewness, and entropy values in RGB, CMYK, RGBA, or CMYKA order (depending on the image type).

@statistics = $image->Statistics();

Finally, the Transform() method accepts a fully-qualified geometry specification for cropping or resizing one or more images. For example,

$p = $image->Transform(crop=>'100x100+0+0');

You can optionally add Image to any method name above. For example, PingImage() is an alias for method Ping().

Handling Exceptions

All PerlMagick methods return an undefined string context upon success. If any problems occur, the error is returned as a string with an embedded numeric status code. A status code less than 400 is a warning. This means that the operation did not complete but was recoverable to some degree. A numeric code greater or equal to 400 is an error and indicates the operation failed completely. Here is how exceptions are returned for the different methods:

Methods which return a number (e.g. Read(), Write()):

$x = $image->Read(...);
warn "$x" if "$x";      # print the error message
$x =~ /(\d+)/;
print $1;               # print the error number
print 0+$x;             # print the number of images read

Methods which operate on an image (e.g. Resize(), Crop()):

$x = $image->Crop(...);
warn "$x" if "$x";      # print the error message
$x =~ /(\d+)/;
print $1;               # print the error number

Methods which return images (EvaluateSequence(), Montage(), Clone()) should be checked for errors this way:

$x = $image->Montage(...);
warn "$x" if !ref($x);  # print the error message
$x =~ /(\d+)/;
print $1;               # print the error number

Here is an example error message:

Error 400: Memory allocation failed

Review the complete list of error and warning codes.

The following illustrates how you can use a numeric status code:

$x = $image->Read('rose.png');
$x =~ /(\d+)/;
die "unable to continue" if ($1 == ResourceLimitError);

Constants

PerlMagick includes these constants:

BlobError
BlobWarning
CacheError
CacheWarning
CoderError
CoderWarning
ConfigureError
ConfigureWarning
CorruptImageError
CorruptImageWarning
DelegateError
DelegateWarning
DrawError
DrawWarning
ErrorException
FatalErrorException
FileOpenError
FileOpenWarning
ImageError
ImageWarning
MissingDelegateError
MissingDelegateWarning
ModuleError
ModuleWarning
Opaque
OptionError
OptionWarning
QuantumDepth
QuantumRange
RegistryError
RegistryWarning
ResourceLimitError
ResourceLimitWarning
StreamError
StreamWarning
Success
Transparent
TypeError
TypeWarning
WarningException
XServerError
XServerWarning

You can access them like this:

Image::Magick->QuantumDepth